(Note: This patent application refers to several publications and patents as indicated with numbers within brackets, e.g., [x]. A list of these publications and patents can be found in the section entitled “References.”)
Gallium nitride (GaN) and its related group-III alloys are key materials for various opto-electronic and electronic devices such as light emitting diodes (LEDs), laser diodes (LDs), microwave power transistors, and solar-blind photo detectors. Currently LEDs are widely used in cell phones, indicators, displays, and LDs are used in data storage disk drives. However, the majority of these devices are grown epitaxially on heterogeneous substrates, such as sapphire and silicon carbide since GaN wafers are extremely expensive compared to these heteroepitaxial substrates. The heteroepitaxial growth of group III-nitride causes highly defected or even cracked films, which hinders the realization of high-end optical and electronic devices, such as high-brightness LEDs for general lighting or high-power microwave transistors.
To solve fundamental problems caused by heteroepitaxy, it is useful to utilize single crystalline group III nitride wafers sliced from bulk group III nitride crystal ingots. For the majority of devices, single crystalline GaN wafers are favorable because it is relatively easy to control the conductivity of the wafer and GaN wafer will provide the smallest lattice/thermal mismatch with device layers. However, due to the high melting point and high nitrogen vapor pressure at elevated temperature, it has been difficult to grow GaN crystal ingots. Growth methods using molten Ga, such as high-pressure high-temperature synthesis [1,2] and sodium flux [3,4], have been proposed to grow GaN crystals, nevertheless the crystal shape grown in molten Ga becomes a thin platelet because molten Ga has low solubility of nitrogen and a low diffusion coefficient of nitrogen.
The ammonothermal method, which is a solvothermal method using high-pressure ammonia as a solvent has demonstrated successful growth of real bulk GaN [5-10]. This new technique is able to grow large GaN crystal ingots, because high-pressure ammonia used as a fluid medium has a high solubility of source materials such as GaN polycrystals or metallic Ga, and high transport speed of dissolved precursors can be achieved. There are mainly three approaches to grow GaN in supercritical ammonia; a method using ammonobasic solutions in single reactor with external heating as disclosed in [6-10] and a method using ammonoacidic solutions in Pt-lined single reactor with external heating as disclosed in [11] and a method using supercritical ammonia with a capsule and internal heaters enclosed in high-pressure reactor as disclosed in [12]. The latter two methods have disadvantages in expanding the reactor scale. For the ammonoacidic approach, it is extremely expensive to use a Pt-liner in a large-scaled pressure vessel. As for the internal capsule, it is structurally very challenging to operate the capsule reactor larger than 2″ diameter. Therefore, the ammonothermal growth using basic mineralizer is the most practical approach to mass-produce bulk GaN. As disclosed in the literature [6, 13-16], GaN has retrograde solubility in supercritical ammonobasic solutions. Therefore, in the conventional ammonothermal growth using basic mineralizer, the temperature for a nutrient zone is set lower than that for a crystallization zone. In addition to this temperature setting, basic ammonothermal method differs in many aspects from other solvothermal methods such as hydrothermal growth of quartz and zinc oxide. Because of this difference, it is not straightforward to apply the solvothermal method to grow group III nitride crystals and more improvements are required to realize mass production of GaN wafers by the ammonothermal method.
First, state-of-the-art ammonothermal method [6-10] lacks scalability into industrial production because it is quite difficult to obtain large enough superalloy material to construct a high-pressure vessel. Since group III nitrides have high melting temperature or dissociation temperature, crystal growth requires relatively higher temperature than other materials grown by the solvothermal method. For example, both quartz and zinc oxide (ZnO) are grown at about 300-400° C. by the hydrothermal method. On the other hand, typical growth temperature of GaN in the ammonothermal method is 450-600° C. [6-10]. Furthermore, our experiment showed that growth at 550° C. or above is typically needed to obtain high-quality crystals. Therefore, Ni—Cr based precipitation hardenable (or age hardenable) superalloy such as Rene-41 (Haynes R-41 or Pyromet 41), Inconel 720 (Pyromet 720), Inconel 718 (Pyromet 718), and Waspaloy A must be used for a vessel material. These superalloys are forged to obtain small-sized, dense grain structure which provides the necessary tensile strength for conditions allowing the solvent to be supercritical. However, if the solid dimension of the piece being worked (such as its thickness) becomes too large, the grain structure necessary for high-tensile strength cannot be obtained by forging. This is because the forging pressure is always applied from the surface during the forging process and the grain size at the inner portion of the material tends to be unaffected if the work size exceeds a certain size. Cracking during the forging/cooling process is also profound for large diameter rods. These problems limit the available size of Ni—Cr based precipitation hardenable superalloys. In case of Rene-41, the maximum available outer diameter for a rod is 12 inch, although the maximum outer diameter for an as-cast (i.e. unforged) rod is larger than 12 inch.
Another obstacle to apply the ammonothermal method to commercial production of GaN single crystals is mediocre quality of grown crystals. Currently, their purity, transparency and structural quality are not sufficient for commercial use. In particular, oxygen concentration is at the order of 1020 cm−3. This high level of oxygen together with Ga vacancy is thought to be the origin of brownish color of GaN grown by the ammonothermal method. The grown crystals also show multiple grains in the growth plane.
Considering above-mentioned limitations, this patent discloses several new ideas to realize a high-pressure vessel that is practically usable for production of group III nitride crystals by the ammonothermal method. This patent also discloses new ways to improve purity, transparency, and structural quality of group III nitride crystals grown by the ammonothermal method.